Protein Engineering, Vol. 13, No. 8, 551-555,
August 2000
© 2000 Oxford University Press
Structural and interactional homology of clinically potential trypsin inhibitors: molecular modelling of Cucurbitaceae family peptides using the X-ray structure of MCTI-II
Department of Biophysics, Bose Institute, P1/2 C.I.T. Scheme VII M, Calcutta-700054, India
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResults and discussionReferences |
|---|
Several trypsin inhibitor peptides (with 28–32 amino acid residues) belonging to the Cucurbitaceae (LA-1, LA-2, MCTI-I, CMTI-I, CMTI-III, CMTI-IV), characterized by a distinctive tertiary fold with three conserved disulphide bonds and with mostly arginine at their active centre, were modelled using the high-resolution X-ray structure of a homologous inhibitor, MCTI-II, isolated from bitter gourd. All the inhibitors were modelled in both their native and complexed state with the trypsin molecule, keeping the active site the same as was observed in the trypsin–MCTI-II complex, by homology modelling using the InsightII program. The minimized energy profile supported the binding constants (binding behaviour) of the inhibitor–trypsin complexes in the solution state. A difference accessible surface area (DASA) study of the trypsin with and without inhibitors revealed the subsites of trypsin where the inhibitors bind. It revealed that the role of mutation of these peptides through evolution is to modulate their inhibitory function depending on the biological need rather than changing the overall structural folding characteristics which are highly conserved. The minor changes of amino acids in the non-conserved regions do not influence significantly the basic conformational and interactional sequences at the trypsin binding subsites during complex formation.
Keywords: Cucurbitaceae/homology modelling/Luffa acutangula/protein/inhibitor complex/trypsin inhibitor
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionReferences |
|---|
The structural and interaction aspects of the small peptides which act as trypsin inhibitors in the different plants of the Cucurbitaceae (squash family) (Hojima et al., 1982
; Wieczorek et al., 1985; Faval et al., 1989; Hara et al., 1989), LA-1, LA-2 from ridged gourd (Luffa acutangula) (Haldar et al., 1996), MCTI-I and MCTI-II from bitter gourd (Momordica charantia) (Huang et al., 1992) and CMTI-I, CMTI-III, CMTI-IV from squash (Cucurbita maxima) (Kupryszewski et al., 1986; Bode et al., 1989) are of growing interest owing to their anti-tryptic activities which can be exploited for the rational design of clinically potent analogues. Bitter gourd (melon) has been widely used in China and South-East Asia for food and to combat immune disorders and common infections (Hayashi et al., 1994). It is capable of inhibiting HIV-1 infection in T-lymphocytes and monocytes and also replication of the virus in infected cells (El-Fiki et al., 1966), and the herb possesses a dose-dependent inhibitory effect on HIV-1 integrase, to impede viral DNA integration and exhibit potent anti-tumour activity against certain human tumour cell lines (Ng et al., 1993). It is useful for therapy of AIDS patients and for herpes virus infections. In our ongoing program in this area, we have already isolated and sequenced LA-1 and LA-2 from Luffa acutangula (Haldar et al., 1996) and are working on a series of trypsin inhibitors of the Cucurbitaceae consisting of 28–32 amino acid residues, with stability over a wide pH range (2–11). In the work presented here, we constructed a series of 3-D models of the trypsin inhibitors of the Cucurbitaceae predicted by homology modelling using the crystal structure of MCTI-II (Huang et al., 1992). Another reason for modelling the 3-D structures of the inhibitors is that no structural information is yet available and the determination of the structures by X-ray and NMR spectroscopy of two of the inhibitors, LA-1 and LA-2 (Luffa acutangula) is in progress, which is time consuming. There is interest in how good the models are before the X-ray and NMR structures are available. Thus the predicted structure will be evaluated in a timely fashion. Structural features predicted from the model that cannot be deduced from sequence alone, such as the substrate-binding sites, will be compared with the forthcoming experimentally determined structure. The involvement of the arginine residue occurs at the reactive/binding centre in all of the inhibitors which constitute the P1 residue of the inhibitor. The sequence homology of these serine protease inhibitors shows about 70% identity with the disulphide bridges at the same position, and with seven totally conserved regions in their primary backbone substantially reveal the homology in basic secondary structural elements, although their cumulative structure–function activities (Creighton and Goldenberg, 1983, 1984) are different. At this interesting stage, we attempted to verify the 3-D structural foldings with interaction studies at the specific binding site of the trypsin molecule (Marquart et al., 1983). Incidentally, the sequence and high-resolution X-ray structure of trypsin complexed with MCTI-II (bitter gourd) have been determined at 1.6 Å resolution (Huang et al., 1992), and it is tempting to deduce other structures in the same family by structural homology modelling from which a rational structure–function relationship can be outlined. Since a difference accessibility of surface area (DASA) (Lee and Richards, 1971) study is useful for structure–activity studies, a critical study of these kinds of inhibitors and inhibitors complexed with trypsin may reflect some aspects of the binding affinity and molecular recognition (Scheidig et al., 1997). Moreover, the substantial interactional sites, hydrogen bonding patterns and geometries of these inhibitors may be assigned through modelling studies in the complexed state with the trypsin molecule and the energy profile of the native and complexed models of these proteins should give an idea of or an approach to a plausible rationale for a better understanding of trypsin–inhibitor complex formation which could be helpful for trypsin–inhibitor design. |
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionReferences |
|---|
The sequences of several peptides belonging to the Cucurbitaceae (Haldar et al., 1996
) with trypsin–inhibitory activities are shown in Table I. These structures were modelled using the coordinates of the inhibitor from the X-ray structure of the trypsin–MCTI-II complex (Huang et al., 1992) obtained from the Brookhaven Protein Data Bank (PDB entry 1MCT). The inhibitors showed about 70% sequence identity. The secondary structure prediction of the inhibitors was performed using the program DSSP (Kabsch and Sander, 1983), July 1995 version. The model building involved (i) sequence alignment of the modelled inhibitors with the template MCTI-II, (ii) replacement of the aligned residues in the structure, (iii) prediction of the three disulphide bond connections according to the spatial relationship of the six cysteines occurring in the sequence, (iv) making insertions by adjusting the backbone conformations of the neighbouring residues and (v) structural refinement with energy minimization. The minimization runs were carried out using the first 500 cycles of the steepest descent method followed by another 500 cycles of the conjugate gradient method. All calculations employed a distance-dependent dielectric and a cut-off for non-bonded terms of 30 Å, in a consistent valence force-field. The minimized structures obtained as described above were subjected to superposition on the X-ray structure of MCTI-II. The root mean square (r.m.s.) difference between the X-ray structure and the modelled minimized structures of the inhibitors was <0.1 Å for the main-chain atoms (Figure 1) and <1.0 Å for all atoms. The refined X-ray structure of trypsin was built using the coordinates from the Brookhaven Protein Data Bank (PDB entry 2PTN) and all the inhibitors were docked into this model, defining an almost identical interaction as found for the trypsin–MCTI-II complex by X-ray studies. The interaction energy between trypsin and the inhibitors, both electrostatic and other forces, were calculated using the Docking module of the InsightII package to find the low-energy (substrate) binding orientation. The energy profile of all the inhibitors, in both the native and complexed state, and the enthalpic component of free energy of binding were calculated (Table II). To optimize trypsin–inhibitor interactions, the minimizations were performed on the side chains of residues which interact to cause inhibition using the conjugate gradient method. The models were subjected to in vacuo energy refinement for 500 steps of conjugate gradient until the r.m.s. derivative of the energy was <0.001. The computer programs InsightII (version 2.2.0) and Homology (Biosym Technologies, San Diego, CA) were used to model 3-D structures and the Discover simulation package (Biosym Technologies) with the consistent valence force-field (Hagler et al., 1985; Dauber-Osguthorpe et al., 1988) was employed for minimization calculations.
|
|
|
Solvent accessibility
To investigate the fit of the contact between trypsin and the inhibitors (MCTI-I, MCT-II, LA-1, LA-2, CMTI-I, CMTI-III, CMTI-IV) (Kupryszewski et al., 1986; Bode et al., 1989; Huang et al., 1992; Haldar et al., 1996), the difference accessibility was calculated and the interacting residues of both trypsin and the inhibitors are shown in Table III. These values are based on the difference between the ASA values of the trypsin structures with and without the inhibitors, where ASA represents the accessible surface area given by Lee and Richards (1971). The trypsin residues 189–195 (Asp, Ser, Cys, Gly, Glu, Ser) interacting with the inhibitor moiety, namely the fifth (Arg) residue, show a decrease in accessibilities, suggesting that these residues of the S1 subsite of trypsin (see Figure 3) are shielded from the solvent phase by binding with the inhibitors and that the large difference accessibility in these regions is due to many interactions with the inhibitor molecule. This difference accessibility value is useful as a parameter in quantitative structure–activity relationship studies based on this complex structure. The specific binding sites in trypsin along with the reactive environment of the inhibitors indicate how binding affinity and molecular recognition (Scheidig et al., 1997) affect the structure–function activity in the system.
|
|
|
Results and discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionReferences |
|---|
In general, six of the trypsin–inhibitor structures amongst the 11 structures given in Table I
belonging to the Cucurbitaceae (LA-1, LA-2, MCT-I, CMTI-I, CMTI-III and CMTI-IV), having seven totally conserved regions along with the disulphide positions, can well be accomplished by homology modelling with the 3-D structure of MCTI-II, which displays no significant structural differences apart from small structural variations at loops and non-conserved regions. As expected, the predicted topology for the aligned regions of the inhibitors is very similar to the X-ray crystal structure of MCTI-II. There are three pairs of disulphide linkages present, C3 and C20, C10 and C22, and C16 and C27. These predictions are based on the spatial relationship between the initial position of the cysteines assigned to the MCTI-II template. However, in addition to the similarity of secondary structural elements and foldings, there are some conformational and sub-interactional differences due to the variation in some amino acid residues occupying equivalent spatial positions. Subsequent docking of the substrates into the trypsin site was employed to identify the major residues that interact with the inhibitors. Several such low-energy binding topologies for trypsin were identified and those most likely to occur showed an identical binding mode of trypsin for both MCTI-II and other inhibitors (Figure 2).
|
The hydrogen bonding characteristics of these complexes are informative. MCTI-II forms eight hydrogen bonds (through the active site arginine), among which the side-chain NH group acts as a hydrogen donor at six centres and main chain carbonyl oxygen acts as an acceptor at two sites. The donor–acceptor behaviour of the reactive site residue (arginine) of the inhibitors also varied; MCTI-I (eight hydrogen bonds of which five are donors and three acceptors), LA-1 (six hydrogen bonds of which three are donors and three acceptors), LA-2 (four hydrogen bonds, of which two are donor and two acceptors), CMTI-IV (eight hydrogen bonds, of which five are donors and three acceptors), CMTI-III (two hydrogen bonds, of which one is a donor and one an acceptor) and CMTI-I (five hydrogen bonds, of which three are donors and two acceptors). In all the inhibitors the arginine at the reactive site (P1 site) interacted through hydrogen bonding at the S1 site of trypsin. Likewise, in LA-2 the arginine at the fifth position interacted whereas the lysine at the eleventh position was not observed to interact (Figure 3
) as reported in a solution study (Haldar et al., 1996). However, these results strongly indicated the potential for interaction of these groups at the above sites during complexation, which may cause a decrease in the free energy of the system and affect the ability to undergo a biochemical reaction via a specific low-energy pathway. The differential donor–acceptor capabilities of the active site residue may lead to variations in the stability of the trypsin–inhibitor complex formed during complexation. The subsequent energy minimization of the active site residues around the docked trypsin leads to some interesting results (Table II) regarding the binding energy of the complex, which can shed light on the interaction affinity of the inhibitors. The decreasing order of enthalpic components of the free energy profile of binding, LA-1 > LA-2 > CMTI-III > MCTI-II > CMTI-I > MCTI-I > CMTI-IV, appears to satisfy the known experimental data on some of the inhibitors in the solution state. Their binding constant values (Ki) decrease in the same order from 120.0 x 10–11 to 0.3 x 10–11. Hence the combinatorial approach of modelling and experimental study helps to increase our understanding of the interactions of the model complex of these inhibitors and indicates the usefulness of modelling in the interpretation of the action of inhibitory agents for designing clinically useful trypsin inhibitors. Ramachandran plot
The majority of the residues of the inhibitors occupy the most favoured regions of the Ramachandran plot and the other residues occupy additional allowed regions as defined in Procheck (Laskowski et al., 1993). No residues of the inhibitors fall in the disallowed region, signifying that the modelled structures are conformationally correct.
Subsites of the protease–inhibitor interface
The S1 subsite of the enzyme and the P1 residue of the inhibitor are the principal components defining the specificity of the enzyme (Figure 3). The S1 binding site of trypsin is a well defined pocket formed by residues 189–195 and 214–220. The main chain and several of the side chains of these segments are configured to make an extensive network of hydrogen bonds with the Arg (P1) residue, which forms the major site of interaction where arginine is the optimal side-chain conformation for tryptic specificity.
The protease interface for the S2' binding cleft of the enzyme is comprised of main-chain atoms from loops 30–41 and 192–193 and by side chains of residues 39 and 40. The important secondary site of interaction, which is the hydrophobic binding cleft, is completed by the side chain of Trp (P2') of the inhibitor residue in MCTI-II. The interface is not a continuum but a set of distinct subsites with different topologies and polar environments.
Conclusion
Most of the serine protease inhibitors from plant sources have been found to have considerable medical and industrial importance and they are being extensively studied to obtain an insight into the protein engineering and their thermostability, design and mechanisms for understanding the specificity of inhibition of enzyme catalysis.
Nature, the supreme biotechnologist, has masterminded extensive engineering of proteins during the process of natural selection in evolution. It is tempting to learn from nature and understand the rules by the careful comparison of convergently/divergently evolved families of protein structure. This can be achieved through the multiple alignment of amino acid sequences via computer processing to identify conserved residues and by the superposition of 3-D structures of family members to identify the conserved signatures of conformations and motifs. The comparison of 3-D structures and knowledge-based homology modelling of trypsin-blocking serine protease inhibitors have shown that functionally significant regions of related proteins can be modelled with high accuracy. The primary purpose of this work was to generate starting models for more precise calculations and comparison with the crystal and NMR structures of the complexes. The derived structures will form a basis for the design of more effective therapeutic agents for the suppression of diseases caused by malfunction/excessive proteolysis. This receptor-based drug design/docking of ligands to receptors appears to satisfy most of the known experimental data. The 3-D structures of these serine protease inhibitors have been elucidated by structural homology modelling using the X-ray structure of MCTI-II in the Protein Data Bank and show that the role of subtle mutations in the sequences of these inhibitors is to modulate delicate specific biological functions and that the overall basic structure is highly conserved and stable.
|
Notes |
|---|
1 To whom correspondence should be addressed. E-mail: ashoke{at}boseinst.ernet.in
|
Acknowledgments |
|---|
We are grateful to Professor N.K. Sinha for the purification protocol to isolate some of the plant inhibitors. We thank DBT, Government of India, for financial support to S.C. and the Director of the Bose Institute for encouragement.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionReferences |
|---|
Bode,W., Greyling,H., Huber,R., Otlewski,J. and Wilusz,T. (1989) FEBS Lett., 242, 285–292.[Web of Science][Medline]
Creighton,T.E. and Goldberg,D.P.O. (1983) Biopolymers, 22, 49–58.[Web of Science][Medline]
Creighton,T.E. and Goldberg,D.P.O. (1984) J. Mol. Biol., 179, 497.[Web of Science][Medline]
Dauber-Osguthorpe,P., Roberts,V.A., Osguthorpe,D.J., Wolff,J., Genest,M. and Hagler,A.T. (1988) Proteins: Struct. Funct. Genet., 4, 31–47.[Web of Science][Medline]
El-Fiki,F.K., Abon-Karma,M.A. and Afify,E.A. (1966) J. Ethnopharmacol., 50, 43–47.
Faval,A., Mattaras,H., Coletti Preveiro,M.A., Zwilling,R., Robinson,E.A. and Castro,B. (1989) Int. J. Pept. Protein Res., 33, 202–208.[Web of Science][Medline]
Hagler,A.T. (1985) In Hruby,V.J. and Meienhofer,J. (eds), The Peptides. Academic Press, New York, pp. 213–299.
Haldar,U., Saha,S., Beavis,R. and Sinha,N. (1996) J. Protein Chem., 15, 177–184.[Web of Science][Medline]
Hara,S., Makino,J. and Ikwnaka,T. (1989) J. Biochem. (Tokyo), 105, 88–92.
Hayashi,K., Takehisa,T., Hamato,N. and Takano,R. (1994) J. Biochem. (Tokyo), 116, 1013–1018.
Hojima,Y., Pierce,V.J. and Pisano,J.J. (1982) Biochemistry, 21, 3741–3746.[Medline]
Huang,Q., Liu,S., Tang,Y., Zeng,F. and Qian,R. (1992) FEBS Lett., 242, 285–292.
Kabsch,W. and Sander,C. (1983) Biopolymers, 22, 2577–2637.[Web of Science][Medline]
Kupryszewski,J., Ragnersons,W., Rolka,K. and Wilusz,T. (1986) Int. J. Pept. Protein Res., 27, 245–250.
Laskowski,R., MacArthur,M., Moss,D. and Thornton,J. (1993) J. Appl. Crystallogr., 26, 91–97.
Lee,B. and Richards,F.M. (1971) J. Mol. Biol., 55, 379–400.[Web of Science][Medline]
Marquart,M., Walter,J., Deisenhofer,J., Bode,W. and Huber,R. (1983) Acta Crystallogr., 20, 306–315.
Ng,T.B., Chang,W.I. and Young,H.W. (1993) Gen. Pharmacol., 24, 655–658.[Web of Science][Medline]
Scheidig,A.J., Hynes,T.R., Pelletier,L.A., Wells,J.A. and Kossaikoff,A. (1997) Protein Sci., 6, 1806–1824.[Web of Science][Medline]
Wieczorek,M., Otlewski,J., Cook,J., Parks,K., Leluk,J., Willimowska-Pele,A., Polanowski,A., Wilusz,T. and Laskowski,M, Jr (1985) Biochem. Biophys. Res. Commun., 126, 646–652.[Web of Science][Medline]
Received October 12, 1999; revised April 29, 2000; accepted May 17, 2000.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  Y.-J. Chen, M.-f. Wu, Y.-h. Yu, M. F. Tam, and T.-Y. Lin Developmental Expression of Three Mungbean Hsc70s and Substrate-binding Specificity of the Encoded Proteins Plant Cell Physiol., November 15, 2004; 45(11): 1603 - 1614. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||